News and events archive

The international classification CSRankings has confirmed the excellence of research at the Faculty of Informatics MU. In the area of Logic & Verification, Masaryk University is ranked 7th in the world, thanks to publications by leading scientists of our faculty. What does this ranking mean and what challenges does the field raise? We talked about this with prof. Antonín Kučera, Head of the Institute of Theoretical Informatics at FI MU and one of the main contributors to the evaluated publications.

Please explain how such an evaluation is created?

The Computer Science Rankings portal compares research institutions according to the number of results published at the world's top conferences in computer science. It divides computer science research into four main areas: AI, Systems, Theory and Interdisciplinary Areas, and a total of 27 sub-areas. For each sub-area, it identifies a list of the most important conferences and then counts the number of papers that authors from relevant institutions publish at those conferences.

Each publication is counted only once, regardless of the number of authors. Input data is obtained from the DBLP database and is updated regularly. The rankings can be compiled for a selected region (e.g. Europe, Czech Republic, or the whole world) and selected subfields of computer science.

How does this assessment reflect on the quality of research institutions?

The obvious advantage of this method of comparison is its objectivity and transparency. It is based on simple, verifiable and difficult to manipulate data. Other relevant indicators, such as journal impact factors, self-assessment reports or the volume of grant funding received, are more problematic from this perspective. Moreover, in computer science, the main measure of the quality of new results is the quality of the conferences at which they are accepted. Unlike other natural sciences, journals typically publish revised and updated versions of conference papers, often several years apart. The predictive value of the data obtained from the Computer Science Rankings is therefore quite high, although of course there is no point in overestimating it. They are still only indicators.

Top 10 institutions in Computer Science: sub-areafield Logic & Verification, comparison: world, period: 2014-2024. Source: CSRangikgs.org (Count = geometric mean count of papers published across all areas; Faculty = number of faculty who have published in these areas)

What falls into the Logic & Verification sub-area in which MU is ranked in the top 10?

This sub-area includes applications of logic in computer science, where the word “logic” should be understood in a broader sense. The focus is on methods for proving the correctness and safety of programs, analysing the long-term behaviour and reliability of computer systems, or designing software that meets given safety criteria. An important new challenge is the safety and reliability of AI, where the knowledge and techniques known from the field of formal verification can be well used and further developed.

Where do you see our faculty specifically excelling?

We have a strong team at the faculty that has long been dedicated to the principles and implementation of tools for program analysis and verification. Almost every year, they win first prizes in various categories of the SV-COMP - International Competition on Software Verification, a worldwide competition that measures the capabilities of formal verification tools on a set of test programs containing specific bugs. The tool that detects the most bugs wins. We also have a long track record of excellent results in the area of probabilistic systems analysis and synthesis, with a significant recent move into AI. We have been able to publish our findings at top AI conferences and as a result we are starting to get more visibility in the Artificial Intelligence subfield. There, however, we are only tied for 243rd out of 524 institutions globally that produce results at this qualitative level. That's not entirely bad, but it's not exactly “dazzling” either. Unfortunately, there is no Czech institution in the top 100.

Where do you see our faculty getting stronger in the future?

Currently, we are already succeeding in transferring new technologies into practice. Some of the faculty's industrial partners are actively using program analysis tools that were developed in direct collaboration with us. At the end of last year, we also established a cooperation with the Regional Military Command in Brno and we are trying to apply our AI technologies in order to increase the effectiveness of critical infrastructure protection. However, in terms of the overall volume of technology transfer, we still have large reserves. There is room for improvement here. 

Prof. RNDr. Antonín Kučera, Ph.D., FI MU

What specifically are you dealing with in your research?

My long-term research interests include analysis of probabilistic systems, automatic synthesis of strategies in stochastic games, and analysis of infinite-state programs. In terms of my personal contribution to the excellent faculty ranking in Logic & Verification, I consider my greatest achievement to be the lion's share of my former PhD students. I am really happy about that.

One of your former PhD students is Prof. Jan Křetínský, who has recently returned to FI thanks to the MUNI Award in Science and Humanities (MASH) grant and is working in the area of AI verification and security. His publications contribute significantly to the success of the faculty in a global comparison. What significance do you think the CSRankings result has within the scientific community?

The popularity of the ranking within the community is growing. At some institutions, the selection process even takes into account how this or that candidate will contribute to moving the institution up within this ranking. Personally, I think that over time CSRankings will gain similar popularity to the Australian CORE database that ranks computer science conferences.

Ranking in the top ten globally in any sub-area of CSRankings is a clear sign of scientific excellence. In the Czech Republic, this is a unique result. However, not to be overly optimistic, the University of Warsaw was ranked 2nd in the Logic & Verification subfield, just behind Oxford. The University of Warsaw is also doing well in many other areas of computer science and is now an unmissable research centre. This can be proved not only by the ranking in CSRankings, but also by the large number of ERC grants received. It is true that their and our starting positions were similar. Here I cannot help but mention that the set-up of the whole system and the way it is managed, both at state and university level, have a part to play in this. In the long term, the differences will inevitably become apparent and the imaginary scissors will continue to open up, which is not good news. Fortunately, we still have excellent students and therefore reasons for optimism.

What do you think would help?

Computer science does not need expensive instrumentation like other natural sciences, except perhaps GPU computing power for artificial intelligence. Thus, the quality of research is a direct reflection of the expertise of the scientists conducting it. To have cutting-edge research in computer science, we also need to be able to recognize it and reward it effectively. There is a great debt here on the part of the state. 

In Poland, they use indicators based on conference rankings to evaluate computer science. Papers at the best conferences are considered to be as valuable as journal publications with an impact factor in the top decile in other sciences. This measure has a direct effect on the flow of money and thus creates a strong incentive to produce top publications. Another difference is that Poland has long applied a career system similar to the American “tenure track” and only truly established scientists who are expected to perform adequately are hired as professors. But on a more positive note: the MASH grants awarded at our university are a great tool for strengthening its professional level, which can be very useful if used properly.

Thank you and we wish you success in your future scientific work.

Author: Marta Vrlová, Office for External Relations and Partnerships at FI MU